Cross-Reaction of Chalcone Synthase and Stilbene Synthase Overexpressed in Escherichia Coli

FEBS 22839 FEBS Letters 460 (1999) 457^461

Cross-reaction of chalcone synthase and stilbene synthase overexpressed in Escherichia coli

Toshio Yamaguchia, Fumiya Kurosakia, Dae-Yeon Suha, Ushio Sankawaa;*, Mizue Nishiokab, Takumi Akiyamab, Masaaki Shibuyab, Yutaka Ebizukab

aFaculty of Pharmaceutical Sciences, Toyama Medical and Pharmaceutical University, 2630 Sugitani, Toyama, 930-0194, Japan bGraduate School of Pharmaceutical Sciences, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan

Received 20 August 1999; received in revised form 4 October 1999

phylogenetic analysis showed a higher homology between Abstract Chalcone synthase (CHS) and stilbene synthase (STS) are related plant polyketide synthases belonging to the CHS and STS from related plants than among CHSs and CHS superfamily. CHS and STS catalyze common condensation STSs, suggesting that STSs have evolved from CHSs several times independently [2]. reactions of p-coumaroyl-CoA and three C2-units from malonyl- CoA but different cyclization reactions to produce naringenin It has been demonstrated that, under certain in vitro con- chalcone and resveratrol, respectively. Using purified Pueraria ditions, CHS produces not only naringenin chalcone but also lobata CHS and Arachis hypogaea STS overexpressed in triacetic acid lactone (TAL) and styrylpyrone as early release Escherichia coli, bisnoryangonin (BNY, the derailed lactone (derailment) byproducts [3,4]. TAL was also detected from the after two condensations) and p-coumaroyltriacetic acid lactone reaction products by fatty acid synthase [5] and certain poly- (the derailed lactone after three condensations) were detected ketide synthases [6,7]. On the other hand, such derailment from the reaction products. More importantly, we found a cross- products have not been observed in the reaction catalyzed reaction between CHS and STS, i.e. resveratrol production by CHS (2.7^4.2% of naringenin) and naringenin production by by STS [8]. STS (1.4^2.3% of resveratrol), possibly due to the conforma- Using puri¢ed Pueraria lobata (kudzu) CHS and Arachis tional flexibility of their active sites. hypogaea (peanut) STS that was overexpressed in Escherichia z 1999 Federation of European Biochemical Societies. coli, we have detected bisnoryangonin (BNY, the derailed lactone after two condensations) and p-coumaroyltriacetic Key words: Chalcone synthase; Stilbene synthase; acid lactone (CTAL, the derailed lactone after three conden- Bisnoryangonin; p-Coumaroyltriacetic acid lactone; sations) from the reaction products produced by either CHS Cross-reaction or STS. Furthermore, we found a cross-reaction between CHS and STS, i.e. resveratrol production by CHS and naringenin production by STS in small but detectable amounts. 1. Introduction 2. Materials and methods Chalcone synthase (CHS, E.C. 2.3.1.74) catalyzes the ¢rst reaction of £avonoid biosynthesis and stilbene synthase (STS, 2.1. Cloning E.C. 2.3.1.95) catalyzes biosynthesis of stilbene backbone of Transcription of STS was induced by cutting peanut seedlings (Sa- stilbene-type phytoalexins. CHS and STS are similar both in kata, Japan) into slices (1 mm thick) that had been immersed in water mechanistic and structural aspects and are representative for 3 days [9]. The slices were then left overnight in water at 25³C in darkness. Total RNA was prepared by the phenol method. cDNA was members of the CHS superfamily [1]. CHS and STS are ho- synthesized using an oligo dT-XbaI primer (5P-GTC GAC TCT AGA modimers of identical subunits of 43 kDa and catalyze repet- (T)15-3P) and PCR-ampli¢ed with primers designed from the published itive decarboxylative condensation of a p-coumaroyl residue A. hypogaea STS sequence [10]; the forward primer (5P-CAT CCA TGG TGT CTG TGA GTG AAT TC-3P) and the reverse primer with three C2-units from malonyl-CoA. Although both enzymes share common starting materials with identical stoichi- (5P-GTA TTA TAT GGC CAT GCT GCG GAG-3P). The forward primer carries an NcoI site suitable for cloning into the expression ometry and apparently use the same condensation mechanism vector, pET-3d. The resulting PCR product was ¢rst blunt-ligated up to the common tetraketide intermediate, CHS and STS into the SmaI site of pBluescriptIISK(+) (MBI Fermentas), rescued catalyze di¡erent ring closure reactions involving di¡erent by NcoI/BamHI digestion, and ligated into the NcoI/BamHI restricted atoms to give rise to di¡erent products, naringenin chalcone vector pET-3d to give pET-3d/STS. The NcoI digestion was carried out in the presence of EtBr (40 Wg/ml) to obtain the 1.2-kb full-length and resveratrol, respectively (Fig. 1). The amino acid homol- cDNA because of the presence of another NcoI site in the coding ogy between these two enzymes is more than 65%, and a region. DNA sequencing was carried out on both strands with several forward and reverse primers using the Dye Terminator Cycle Se- quencing kit from Applied Biosystems (CA, USA). Construction of pET-3d/CHS using similar methods has been described [11].

*Corresponding author. Fax: (81) (76) 434-5170. 2.2. Protein expression and puri¢cation E-mail: [email protected] E. coli BL21(DE3)pLysE cells transformed with either pET-3d/CHS or pET-3d/STS were grown at 37³C in LB medium containing 50 Wg/ Abbreviations: CHS, chalcone synthase; STS, stilbene synthase; ml ampicillin and 25 Wg/ml chloramphenicol and induced with 0.4 mM BNY, bisnoryangonin; CTAL, p-coumaroyltriacetic acid lactone; IPTG. After an induction period of 4 h at 37³C, the cells were har- TAL, triacetic acid lactone; KPi, potassium phosphate; Trx-, thiore- vested, washed and suspended in bu¡er A (10 mM potassium phos- doxin- phate (KPi) bu¡er (pH 7.5), 2 mM DTT). Following sonication (10 s

FEBS 22839 28-10-99 458 T. Yamaguchi et al./FEBS Letters 460 (1999) 457^461

U10) on ice and centrifugation (12 000Ug, 4³C, 15 min), soluble obtained were redissolved in 500 Wl of 50% aqueous methanol con- fractions were obtained. After (NH4)2SO4 precipitation (30^80% sat- taining 1% acetic acid in which authentic resveratrol (1.5 Wg) and uration), CHS was puri¢ed to apparent homogeneity by successive naringenin (2 Wg) were added. Using a Hitachi 655 HPLC system chromatography procedures: (1) DEAE-Toyopearl (TOSOH, Tokyo, and ODS-80Ts column (TOSOH), the radioactive products were sep- Japan) anion exchange chromatography using a linear gradient of arated with isocratic elution with 50% aqueous methanol containing NaCl (0^500 mM) in bu¡er A, (2) Bio-Gel hydroxyapatite (Bio- 1% acetic acid for 30 min. Radioactivity of each 0.5 ml fraction was Rad) using a linear gradient of 10^300 mM KPi (pH 7.5) containing measured with a L-scintillation counter and the fractions correspond- 2 mM DTT and (3) chromatofocusing on polybu¡er exchanger ing to main product (naringenin for CHS and resveratrol for STS) (PBE94, Pharmacia) pre-equilibrated with 25 mM imidazole-HCl and cross-reaction product (resveratrol for CHS and naringenin for (pH 7.4), 2 mM DTT and eluted with Polybu¡er74-HCl (pH 5.0), STS) were collected. For carrier dilution analysis, recrystallization of 2 mM DTT. STS was similarly puri¢ed after successive DEAE ion the radioactive products with carrier compounds was performed at exchange and hydroapatite chromatography steps. Protein contents least four times in two di¡erent solvent systems (methanol and ethyl were determined with Bradford's dye method (Bio-Rad) [12], using acetate/hexane) and speci¢c radioactivity was measured after each bovine serum albumin as standard. recrystallization step. Expression and puri¢cation of P. lobata CHS and A. hypogaea STS The identity of the cross-reaction products was further con¢rmed as thioredoxin-HisTag fusion proteins (Trx-CHS and Trx-STS) used by GC-MS and LC-MS. First, the CHS and STS reaction products in the LC-MS analysis of cross-reaction will be described elsewhere. were methylated with trimethylsilyldiazomethane and analyzed by GC-MS as described above. From the methylated CHS reaction 2.3. Enzyme assay products, a peak of m/z = 256 (Rt = 12.99 min) corresponding to di- For both CHS and STS, the reaction mixture (0.1 ml) contained methylresveratrol was detected with mono- (m/z 286), di- (m/z 300) appropriate amounts of protein (2^5 Wg of puri¢ed enzyme), 0.1 mM and trimethylnaringenin (m/z 314). Likewise, a peak of m/z = 300 14 p-coumaroyl-CoA and 16.8 WM [2- C]malonyl-CoA (2.2 GBq/mmol, (Rt = 15.35 min) corresponding to dimethylnaringenin was detected NEN) in 0.1 M KPi bu¡er (pH 7.5) (K2HPO4/KH2PO4) containing from the methylated STS reaction mixture together with di- (m/z 2 mM DTT. After incubation at 37³C for 30 min, 50 Wl of 50% acetic 256) and trimethylresveratrol (m/z 270). For LC-MS analysis, Trx- acid was added and the reaction products were extracted with 200 Wl CHS (11 mg, 250 pkat) or Trx-STS (9.5 mg, 160 pkat) was incubated of ethyl acetate. A portion (50 Wl) of the extract was analyzed by TLC with 0.1 mM p-coumaroyl-CoA and 0.3 mM malonyl-CoA in 50 ml of on RP18 plate (Merck) with methanol:H2O:acetic acid (60:40:1) as 0.1 M KPi bu¡er (pH 7.4) containing 1 mM DTT for 40 min at 37³C. solvent and the radioactive products were quanti¢ed with an Imaging At the end of incubation, the pH was raised to 9 with 1 N NaOH, and Plate analyzer (BAS2000, Fuji). Authentic compounds, TAL the reaction mixture was further incubated at 37³C for another 10 min (Rf = 0.65, Tokyo Kasei), resveratrol (Rf = 0.50, Sigma) and naringe- to complete chemical transformation of naringenin chalcone to nar- nin (Rf = 0.32, Nacalai tesque) were used as internal standards to ingenin [13]. The ethyl acetate extract was dried over Na2SO4 and the identify the products. The speci¢c enzyme activity was expressed in solvent was removed under vacuum at 30³C. The remaining products pmol of the product produced/s/mg (pkat/mg). were analyzed by LC-MS (Thermoquest LCQ) as previously described [14]. 2.4. Identi¢cation of bisnoryangonin and p-coumaroyltriacetic acid Resveratrol: LC-APCIMS (positive), 229.1 [M+H], MS/MS (pre- lactone cursor ion at 229) m/z (rel. %) 211 (65), 135 (100), 119 (12), 107 (11). For identi¢cation of BNY, the ethyl acetate-extracted products Naringenin: LC-APCIMS (positive), 273.2 [M+H], MS/MS (precur- produced by STS and CHS were ¢rst separated by HPLC. The col- sor ion at 273) m/z (rel. %) 171 (18), 153 (80), 147 (100). umn (5 Wm, 4.6 mm I.D. U150 mm, ODS80Ts, TOSOH) was washed at a £ow rate of 0.6 ml/min at 40³C with 45% methanol in H2O containing 1% acetic acid for 10 min and then the concentration of 3. Results methanol was increased to 55% during the next 30 min. Detection was with an UV monitor at 254 nm. The peak at Rt = 16.3 min was collected and, after evaporation of the solvent, the residues were dis- 3.1. Sequence of A. hypogaea STS solved in methanol and reacted with trimethylsilyldiazomethane in the When the deduced amino acid sequence of Japanese peanut dark for 30 min at 22³C. The resulting sample was subjected to GC- STS cloned in this study (Accession number: AB027606) was MS (Hewlett Packard 5890 coupled with JEOL JMS-SX102A) analysis using a DB5 column (0.25 mm I.D. U30 m, ¢lm thickness 0.25 compared to the published peanut STS sequence (Swiss-Prot Wm, J and W) at 200³C to 250³C (10³C/min). The retention time and P20178) [10], 8 out of 389 amino acids were found to be mass spectrum were found to be identical to those of authentic yan- di¡erent. Those are Ser-46 corresponding to Gly-46 in Ref. gonin: GC, Rt = 22 min; EIMS (70 eV, m/z), 258 (M ), 230, 187. [10], Ser-195 to Ala-195, Ser-202 to Asn-202, Leu-236 to Ile- Production of CTAL by CHS and STS was con¢rmed by LC-MS 236, Asp-281 to Gly-281, Pro-304 to Leu-304, Glu-315 to (see below). CTAL: Multi-stage LC-APCIMS, MS m/z 273.1 [M+H], 271.1 [M3H]3, MS/MS (precursor ion of m/z 271.1) 227.2 Gln-315 and Ala-356 to Thr-356. This di¡erence in amino 3 [M3H3CO2] , MS/MS/MS (transitions m/z 271.1 s 227.2) 185.2 acid sequence probably re£ects di¡erent origins of the plant 3 [M3H3CO23CH2CO] . materials used.

2.5. Cross-reaction The enzyme reaction was carried out under the same conditions as 3.2. Puri¢cation of CHS and STS described above except that the reaction volume was 1 ml and about Puri¢cation of CHS and STS overexpressed in E. coli is 50 Wg of enzyme (10^40 pkat/mg) was used. Radioactive products thus summarized in Table 1. Typical experiments yielded 0.3^0.4

Table 1 Puri¢cation of P. lobata CHS and A. hypogaea STS overexpressed in E. coli Stage Total protein Total activity Speci¢c activity Recovery Puri¢cation (mg) (pkat) (pkat/mg) (%) (fold) CHS Crude extract 29 36 1.2 (100) ^ DEAE ion exchange 4.5 9.1 2.0 25 1.7 Hydroxyapatite 1.7 6.5 3.8 18 3.2 Chromatofocusing 0.42 4.2 10 12 8.3 STS Crude extract 10 35 3.5 (100) ^ DEAE ion exchange 1.3 14 10 39 2.9 Hydroxyapatite 0.31 12 37 33 11

FEBS 22839 28-10-99 T. Yamaguchi et al./FEBS Letters 460 (1999) 457^461 459

are present in the reaction mixture as free acids at neutral pH, thus poorly extractable by ethyl acetate, and that lactonization to form pyrone ring is non-enzymatic. Similar pH-de- pendence of extraction of pyrone derivatives was also re- ported recently by others [15]. Whereas CTAL was a major byproduct of the CHS reaction, BNY was a major byproduct of the STS reaction (Fig. 2B).

3.4. Cross-reaction When the reaction products produced by puri¢ed recombinant enzyme were analyzed by HPLC, a small radioactive peak corresponding to resveratrol (Rt = 10 min) was detected in the CHS products and a peak corresponding to naringenin (Rt = 20 min) was detected in the STS products (Fig. 3). Cross-reaction was veri¢ed by carrier dilution analysis. Suc- cessive rounds of recrystallization of radioactive products in the presence of non-labeled compounds gave crystals of con- stant speci¢c radioactivity. For resveratrol produced by CHS, the speci¢c radioactivity was 8200 dpm/mmol (methanol), 7300 (methanol), 7000 (ethyl acetate/hexane) and 6700 (ethyl acetate/hexane), and for naringenin produced by STS, it was 8900 dpm/mmol (methanol), 7900 (methanol), 7500 (ethyl acetate/hexane) and 8300 (ethyl acetate/hexane). The cross-reaction was determined to be 2.7^4.2% (resveratrol/naringenin) for CHS and 1.4^2.3% (naringenin/resveratrol) for STS. The cross-reaction was further con¢rmed by GC-MS and LC-MS analyses. The retention time, molecular mass and major Fig. 1. Reaction steps catalyzed by CHS and STS and the products daughter ions in MS/MS were indistinguishable to those of produced in vitro. authentic compounds and those found in the literature [16]. Trx fusion enzymes were indistinguishable from the native enzymes with regard to the pattern of byproduct production mg of apparently homogeneous enzymes from 300 ml bacteria and kinetic parameters (data not shown). culture (Fig. 2A). Recovery was 10^30% with about 10-fold enrichment.

3.3. Identi¢cation of BNY and p-coumaroyltriacetic acid lactone Production of BNY and CTAL by STS as well as by CHS was clearly visualized by TLC (Fig. 2B). The TLC spots were identi¢ed with non-labeled carrier compounds (resveratrol and naringenin) and with parallel experiment using p-coumaroyltriacetic acid synthase which has been recently cloned from Hydrangea macrophylla and shown to produce CTAL and BNY as the main product and a major byproduct, respectively [14]. Further, the identity of BNY and CTAL was unambig- uously determined by GC-MS and LC-MS. The production of these derailment products by CHS and STS were dependent on the presence of sulfhydryl reagents such as DTT in the reaction mixture, consistent with the earlier report with CHS [3]. It was thought that sulfhydryl reagent would act as an acceptor to the enzyme-bound triketide (or tetraketide) intermediate in a non-enzymatic trans-esteri¢cation to form unstable thiol adduct, which should be immediately converted to the corresponding pyrone [3]. In agreement with this no- Fig. 2. Puri¢cation and enzyme assay of P. lobata CHS and A. hy- tion, DTT showed apparently parallel e¡ects on formation of pogaea STS. A: SDS-PAGE was performed on a 12% acrylamide minislab gel under reducing conditions. For both CHS and STS, BNY and CTAL. However, the requirement of the sulfhydryl lane 1 contained 3 Wg of proteins obtained after (NH4)2SO4 frac- reagent was not absolute, as these byproducts were still de- tionation and lane 2 contained 1 Wg of puri¢ed enzyme. The pro- tectable when the enzyme reaction was carried out in the teins were stained with Coomassie blue R250 (Bio-Rad). The num- absence of DTT (data not shown). Acidi¢cation of the reac- bers on left indicate molecular weight in kDa of the Mr marker proteins. B: Radio thin layer chromatogram of enzyme assays using tion mixture to pH 6 5 greatly enhanced the amounts of BNY puri¢ed recombinant CHS (5 Wg) and STS (2 Wg) and p-coumaroyl- and CTAL detected on TLC. These results suggest that the CoA and [2-14C]malonyl-CoA as substrates. The conditions for en- majority of these compounds, both being pyrone derivatives, zyme assay and TLC analysis are described in Section 2.

FEBS 22839 28-10-99 460 T. Yamaguchi et al./FEBS Letters 460 (1999) 457^461

STS produce counterpart products in low amounts (2^4% of main products) which could not be detected by routine assays. The possibility that the cross-reaction products were formed during work-up by chemical transformation was thought to be unlikely based on the following. First, although the corresponding chemical transformations are known, they are nor- mally carried out under harsh conditions unlike the work-up procedures used in this study. For example, the Dieckmann condensation of methyl ester of 7-phenyl-3,5,7-trioxohepta- noic acid to corresponding acylphloroglucinol required a pro- longed incubation under strongly basic conditions, such as 2 M KOH [18]. Even though aldol condensation of triketo acids can proceed in weakly acidic (pH 5) aqueous bu¡er, the products are fairly stable L-resorcylic acids [18]. Decarboxyl- ation of aromatic carboxylic acids such as conversion of 6-styryl-L-resorcylic acid to pinosylvin and of orsellinic acid to orcinol have been achieved either by heat (180³C) [18,19], by alcoholic KOH [20] or by a decarboxylase [21]. It should be noted that stilbene carboxylic acids, though found in nature [22], have not been detected from the STS reaction mixture. Secondly, neither naringenin nor resveratrol was detected after using an identical work-up procedure in the reaction mixture produced by p-coumaroyltriacetic acid synthase (CTAS), which catalyzes, like CHS and STS, three condensation reactions between p-coumaroyl-CoA and malonyl-CoA to triketo acid but does not catalyze any cyclization reaction [14]. This suggested that lactonization of the free triketo acid Fig. 3. HPLC analysis of reaction products produced by puri¢ed recombinant P. lobata CHS (A) and A. hypogaea STS (B). Authentic predominates during work-up to give CTAL as a major by- resveratrol and naringenin were added to the reaction products to product (Fig. 1) and, therefore, we concluded that the cross- aid detection by UV absorbance (thin line) and the resulting sample reaction of CHS and STS was most likely enzymatic of na- was injected onto a TOSOH ODS80-Ts column. The products were ture. separated with an isocratic elution of 50% aqueous methanol con- One reasonable interpretation for this seemingly unnatural taining 1% acetic acid. Radioactivity (thick line) of each fraction (0.5 ml) was measured and fractions corresponding to the cross-re- cross-reaction is that parts of heterologously overexpressed action products (resveratrol for CHS and naringenin for STS) were enzymes are in unnatural conformational states and these pooled and used for carrier dilution analysis. misfolded, yet soluble enzymes are responsible for the observed cross-reaction. Overexpression in E. coli often leads 4. Discussion to misfolding especially for membrane-associated eukaryotic proteins, as bacteria lack an intracellular membrane environ- CHS puri¢ed from plants was shown to produce styrylpyr- ment that may be necessary for e¤cient and accurate folding ones and other minor byproducts in certain in vitro conditions [23]. Both CHS and STS are shown to be weakly membrane- [3,4]. In the present study, we showed that, like CHS, puri¢ed associated in the plant cells [24,25] and CHS is associated with recombinant peanut STS produced not only BNY, a styryl- other enzymes involved in the £avonoid biosynthesis pathway pyrone, but also CTAL, the recently found derailment prod- [17]. However, an attempt to resolve conformational hetero- uct at the stage of tetraketide intermediate [14]. This was not geneity using hydroxyapatite chromatography was unsuccess- unexpected since it has been thought that CHS and STS cata- ful, as cross-reaction activities were co-puri¢ed with the nat- lyze identical condensation steps leading to the linear tetrake- ural activities (data not shown). Another intriguing and tide intermediate prior to di¡erent cyclization steps [1]. De- probably more conceivable possibility is that CHS and STS tection of BNY and CTAL from the STS reaction products possess intrinsic potential to catalyze cross-reaction possibly provides an unequivocal evidence for this notion. Under the due to conformational £exibility of their active sites, and that conditions employed, CHS produces more byproducts than this intrinsic, yet `cryptic' capability of these enzymes to per- STS (Figs. 2B and 3). This may re£ect the more loosely struc- form the cross-reaction was manifested under non-optimal tured active site of CHS compared with that of STS, in vitro, folding conditions during bacterial overexpression. Our recent which allows to a greater extent premature thioester exchange ¢nding that cross-reaction of CHS and STS expressed in eu- with sulfhydryl reagent present in the reaction mixture or karyotic systems (yeasts and insect cells) apparently decreased thioester hydrolysis by water to give derailment products. as compared to the cross-reaction found in this study with CHS is believed to be present in the cell as a part of multi- bacterial expression system seems to support this idea (San- enzyme complex of £avonoid biosynthesis including polyke- kawa, unpublished data). In line with this idea, it should be tide reductase, chalcone isomerase and other modifying en- noted that cyclization steps in CHS and STS are proposed to zymes to achieve `metabolite channeling' [1,17]. be under pure topological control [8]. Indeed, the recently Cyclization steps in CHS and STS are di¡erent; Dieckmann solved crystal structure of alfalfa (Medicago sativa) CHS by condensation leading to naringenin chalcone in CHS and al- Ferrer et al. did not reveal apparent functional amino acids in dol condensation to stilbene in STS. Nonetheless, CHS and the cyclization pocket that might participate in a speci¢c

FEBS 22839 28-10-99 T. Yamaguchi et al./FEBS Letters 460 (1999) 457^461 461

CHS-type cyclization reaction [26]. The authors, therefore, [6] Spencer, J.B. and Jordan, P.M. (1992) Biochem. J. 288, 839^ proposed that alterations in the surface topology of the cyc- 846. [7] Kurosaki, F., Kizawa, Y. and Nishi, A. (1989) Eur. J. Biochem. lization pockets of CHS and STS may a¡ect the stereochem- 185, 85^89. istry of the cyclization reactions. [8] Scho«ppner, A. and Kindl, H. (1984) J. Biol. Chem. 259, 6806^ Based on phylogenetic analysis, Schro«der et al. [2] proposed 6811. di¡erent STSs evolved from CHSs independently several times [9] Aguamah, G.E., Langcake, P., Leworthy, D.P., Page, J.A., Pryce, R.J. and Strange, R.N. (1981) Phytochemistry 20, 1381^ by a limited number of amino acid exchanges. However, con- 1383. sensus (signature) amino acid sequences characteristic to CHS [10] Schro«der, G., Brown, J.W.S. and Schro«der, J. (1988) Eur. J. or to STS have yet to be recognized. It may result from the Biochem. 172, 161^169. insu¤cient number of STS sequences available (6 entries in [11] Sankawa, U. and Hakamatsuka, T. (1997) in: Dynamic Aspects Swiss-Prot). On the other hand, STS conceivably gained its of Natural Products Chemistry (Ogura, K. and Sankawa, U., Eds.), pp. 25^48, Kodansha Scienti¢c, Tokyo. new activity by changing active site geometry rather than by [12] Bradford, M.M. (1976) Anal. Biochem. 72, 248^254. introducing a few new functional amino acids in the active site [13] Tropf, S., Ka«rcher, B., Schro«der, G. and Schro«der, J. (1995) of STS. A closely related case can be found in chemical diver- J. Biol. Chem. 270, 7922^7928. sity of higher plant fatty acids, where independent, multiple [14] Akiyama, T., Shibuya, M., Liu, H.-M. and Ebizuka, Y. (1999) Eur. J. Biochem. 263, 834^839. evolution of oleate 12-hydroxylase from oleate 12-desaturase [15] Zuurbier, K.W.M., Leser, J., Berger, T., Hofte, A.J.P., Schro«der, was proposed. Site-directed mutagenesis studies involving as G., Verpoorte, R. and Schro«der, J. (1998) Phytochemistry 49, few as seven amino acids suggested that catalytic plasticity in 1945^1951. these two enzymes is achieved by changing active site geom- [16] Weintraub, R.A., Ameer, B., Johnson, J.V. and Yost, R.A. etry leading to di¡erent partitioning of a common intermedi- (1995) J. Agric. Food Chem. 43, 1965^1968. [17] Hrazdina, G. and Wagner, G.J. (1985) Arch. Biochem. Biophys. ate into di¡erent reaction products [27]. Clearly, further stud- 237, 88^100. ies with regard to the origin of the cross-reaction of CHS and [18] Harris, T.M. and Carney, R.L. (1967) J. Am. Chem. Soc. 89, STS overexpressed in E. coli are warranted. 6734^6741. [19] Harris, T.M. and Harris, C.M. (1986) Pure Appl. Chem. 58, 283^ Acknowledgements: The authors are grateful to the Ministry of Edu- 294. cation, Science, Sports and Culture of Japan for Grants-in Aid for [20] Bentley, R. and Zwitkowits, P.M. (1967) J. Am. Chem. Soc. 89, Scienti¢c Research (B) (No. 09044212) and (C) (No. 10680564). D.-Y. 676^680. Suh thanks the Tokyo Biochemistry Research Foundation for a fel- [21] Mosbach, K. and Schultz, J. (1971) Eur. J. Biochem. 22, 485^ lowship (TBRF-98-10). 488. [22] Yoshikawa, M., Matsuda, H., Shimoda, H., Shimada, H., Har- ada, E., Naitoh, Y., Miki, A., Yamahara, J. and Murakami, N. References (1996) Chem. Pharm. Bull. 44, 1440^1447. [23] Gonzales, F.J., Kimura, S., Tamura, S. and Gelboin, H.V. (1991) [1] Schro«der, J. (1999) in: Comprehensive Natural Products Chem- Methods Enzymol. 206, 93^99. istry (Sankawa, U., Ed), Vol. 1, pp. 749^771, Elsevier, Amster- [24] Rupprich, N. and Kindl, H. (1978) Hoppe-Seyler's Z. Physiol. dam. Chem. 359, 165^172. [2] Tropf, S., Lanz, T., Rensing, S.A., Schro«der, J. and Schro«der, G. [25] Hrazdina, G., Zobel, A.M. and Hoch, H.C. (1987) Proc. Natl. (1994) J. Mol. Evol. 38, 610^618. Acad. Sci. USA 84, 8966^8970. [3] Kreuzaler, F. and Hahlbrock, K. (1975) Arch. Biochem. Biophys. [26] Ferrer, J.-L., Jez, J.M., Bowman, M.E., Dixon, R.A. and Noel, 169, 84^90. J.P. (1999) Nat. Struct. Biol. 6, 775^784. [4] Hrazdina, G., Kreuzaler, F., Hahlbrock, K. and Grisebach, H. [27] Broun, P., Shanklin, J., Whittle, E. and Somerville, C. (1998) (1976) Arch. Biochem. Biophys. 175, 392^399. Science 282, 1315^1317. [5] Yalpani, M., Willecke, K. and Lynen, F. (1969) Eur. J. Biochem. 8, 495^502.

FEBS 22839 28-10-99